Use content length to avoid http hang

[bmidgley]

response.content() is hanging on responses that don't end with a final newline.

Issue is in #546

Here's my actual workaround while waiting for the real fix... reach into the impl and read one character at a time. Stop when reaching "}"

    response = urequests.get(url)
    response._cached = response.raw.read(1)
    while True:
        response._cached += response.raw.read(1)
        if response._cached[-1] == 125: break
    json = response.json()

[bmidgley]

Is there another way to avoid this hang?

[andrewleech]

I don't think the .read() function should be waiting for a newline, only .readline() is supposed to work this way.
.read() should just provide all data currently available.
The existing implementation can work for downloading binary files too, which definitely doesn't depend on a trailing newline.

That being said it does seem sensible to me to honour the content length header if it's provided.

[dpgeorge]

Closing in favour of #1124, which implements HTTP/1.1.